If basing conclusions on impaired hearing among school-children in Sisimiut, the prevalence varies from 17% if based on self-reports, to 10% if based on clinical examination from the threshold chosen in the present survey. Choice of threshold levels determines the variation between the two types of measurement, and there are large inconsistencies identified when trying to associate the two measures.
A main finding of the present analyses is, that the traditional measures of correlation are best suited to illustrate that the two measures are not closely related. Kappa coefficient is used to measure inter-rater reliability, and taking account the possibility of an agreement occurring by chance. However, the underlying assumption of using Kappa is that the two measures are measuring the same, an assumption that must be rejected to hold.
With this finding, it becomes evident, that the clinical measures cannot be considered as an exact measure, and that researchers and clinicians should allow for a “grey zone” of classifications including the personal assessment of hearing impairment. This “grey zone” can be quantified by further analyses of the 30 out of 185 children that showed inconsistencies between the SR and CM, that were first evaluated for a second opinion, and initial findings were confirmed. Given the degree of variations, the best conclusion is to consider the two measurements of hearing impairment as covering substantially different areas, and the explanation could be found in that the one is a consequence of the other. This is supported by a strong statistical association of SR by CM, with an OR of 8.2.
As described by Schnohr et al. (2019) the item measuring self-reported hearing impairment was associated to the school-situation, since the item was developed for school-aged children. In the event, that a respondent was asked why he/she did not keep up in school, and the suggestion was made that is was due to impaired hearing, such a question can simply nudge the respondent to reply ‘yes’ even though the underlying cause – to not paying attention to school – is something else. Often children are asked to pay attention for other reasons than impaired hearing, and it would be relevant to adjust for those related causes. Age, gender, socioeconomic background and overall self-perceived health has been demonstrated to be associated with disagreement between self-reported hearing and audiometric outcome among older adults and may also affect the responses of children (15;16). However, in a sample size of 185 respondents, it is not statistically sound to examine associations of related predictors on impaired hearing, to e.g. study the school environment and other aspects of the respondent’s health, but it would be interesting and relevant for future studies.
Important factors that could lead to self-reported hearing difficulties in school but a normal pure-tone audiogram at examination are auditory neuropathy spectrum disorders, auditory processing disorders and any attention deficit disorder. Furthermore, the acoustic environment in the school is important. Background noise in the classroom, size of the room, number and position of students will affect their ability to hear the teachers voice
In Greenland it is only possible – as in this study – to use pure tone audiometry. Hearing is a highly complex neurological process that involve not only the outer and inner ear but also lower and higher cerebral functions that are interacting and are influenced by psychological processes and by the broader context in which the hearing takes place. The individual perception of what is heard therefore is a meta process.
Considerations can be taken to what the measurement properties are for each of the two methods, and taking into account if the conclusions derived should stem from a representative result for the entire population – in which case an epidemiological survey of a large representative samples of the population is most appropriate – or whether researchers are looking for clinical conditions causing hearing impairment – in which case clinical examinations of high risk populations is most appropriate.
Among older populations audiometric measures was found to explain less than 50% of the variance in hearing handicap and it was suggested that hearing handicap in the elderly will be measured more appropriate via a self-report format rather than as an inference from audiometric data (17)
In spite of clear findings that the two measurements are not measuring the same construct, it is clear that they are closely related. Irrespectively where the threshold levels are set, it is fair to presume that students that are classified as normal hearing will have problems following school, and it becomes evident that both measures have each their relevance, depending on the purpose and method of any given investigation. Based on the present study, it is not appropriate to conclude whether the self-reported or the measured collected from clinical examinations have the highest reliability, as both measurements have uncertainties, and there is no golden standard available at present.
In any case, the Arctic population is a high risk population for chronic conditions related to impaired hearing, and a continued focus on the monitoring and examinations of both representative and high-risk populations serves as an important topic for future public health in Greenland.